- Email: [email protected]
Expression Pattern Analysis of Hepatocellular Carcinoma Tumor Markers in Viral Hepatitis B and C Patients Undergoing Liver Transplantation and Resection
Expression Pattern Analysis of Hepatocellular Carcinoma Tumor Markers in Viral Hepatitis B and C Patients Undergoing Liver Transplantation and Resection T.Y. Ha, S. Hwang, K.H. Kim, Y.J. Lee, C.S. Ahn, D.B. Moon, G.W. Song, K.M. Park, N. Kim, and S.G. Lee ABSTRACT Background. This study was conducted to compare the expression patterns of serum alpha-fetoprotein (AFP) and proteins induced by vitamin K absence or antagonist-II (PIVKA-II) in hepatocellular carcinoma (HCC) patients undergoing liver transplantation (LT) and resection at a high-volume single institution. Methods. First, 663 liver transplant recipients with HCC were selected. They were divided into hepatitis B virus (HBV) (n ¼ 628) and hepatitis C virus (HCV) groups (n ¼ 35). Their medical records were retrospectively reviewed. Second, another cohort of 2709 patients who underwent HCC resection included 2258 HBV, 143 HCV, and 308 non-HBV non-HCV (NBNC) patients. Results. In the transplantation group, pretransplantation AFP level >20 ng/mL was observed in 42.5% of HBV patients and 60% of HCV patients (P ¼ .042). PIVKA-II level >40 mAU/mL was observed in 30.6% of HBV patients and 42.9% of HCV patients (P ¼ .127). In the resection group, a preoperative AFP level >20 ng/mL was observed in 51.7% of HBV patients and 43.3% of HCV patients (P ¼ .052). PIVKA-II level >40 mAU/mL was observed in 59.7% of HBV patients and 56.6% of HCV patients (P ¼ .47). Preoperative AFP level >20 ng/mL and PIVKA-II level >40 mAU/mL were observed in 35.7% and 61% of NBNC patients, respectively. Receiver-operator characteristic curve analyses revealed that the expression pattern of PIVKA-II in patients with elevated AFP level was not predictable and vice versa, regardless of background liver diseases. Conclusions. This study indicates that serum AFP and PIVKA-II may be expressed variably regardless of the types of background liver disease. Further large-volume multicenter studies are needed to evaluate the possibility of the etiology-dependent expression of tumor markers.
I
T IS controversial that different tumor markers are often emphasized between Korea and Japan for the diagnosis and follow-up treatment of hepatocellular carcinoma (HCC) in liver resection patients and liver transplant recipients. In fact, alpha-fetoprotein (AFP) seems to be more frequently emphasized than protein induced by vitamin K absence or antagonist-II (PIVKA-II or des-gamma-carboxy prothrombin [DCP]) in Korea, where hepatitis B virus (HBV) infection is predominant, and vice versa in Japan, where hepatitis C virus (HCV) infection is more prevalent [1e6]. At this time, a question was raised regarding the possibility of a different expression of HCC tumor marker according to the primary diseases of the background livers. This study was conducted to compare the expression 0041-1345/14/$esee front matter http://dx.doi.org/10.1016/j.transproceed.2013.12.037 888
patterns of serum AFP and PIVKA-II in HCC patients undergoing liver transplantation and liver resection at a high-volume single institution.
From the Department of Surgery (S.H., K.H.K., Y.J.L., C.S.A., D.B.M., T.Y.H., G.W.S., K.M.P., S.G.L.) and the Asan Institute for Life Sciences (N.K.), Asan Medical Center, University of Ulsan College of Medicine, Seoul, Korea. Address reprint requests to Shin Hwang, MD, PhD, Division of Hepatobiliary Surgery and Liver Transplantation, Department of Surgery, Asan Medical Center, University of Ulsan College of Medicine, Poongnap-dong, Songpa-gu, Seoul, 138-736, Korea. E-mail: [email protected] ª 2014 by Elsevier Inc. All rights reserved. 360 Park Avenue South, New York, NY 10010-1710 Transplantation Proceedings, 46, 888e893 (2014)
HEPATOCELLULAR CARCINOMA TUMOR MARKERS
889
PATIENTS AND METHODS In part 1 of this study, using our institution’s liver transplantation database containing more than 3000 adult cases, 663 recipients who had been diagnosed with viral hepatitis and HCC were selected. To objectively compare the patient parameters, patients with AFP <1000 ng/mL and PIVKA-II <1000 mAU/mL were included (Table 1). They were divided into the HBV group (n ¼ 628) and the HCV group (n ¼ 35), and their medical records were retrospectively reviewed. In part 2 of this study, a cohort of 2709 patients who underwent HCC resection, which consisted of consecutive 2258 HBV patients, 143 HCV patients, and 308 non-HBV non-HCV (NBNC) patients, had their AFP and PIVKA-II expression levels analyzed (Table 2). If both AFP and PIVKA-II were not measured at the same time, such patients were excluded from this study. No value limitation for exclusion was applied to the HCC resection group. Our study protocol was approved by the institutional review board of our institution. All numerical data are presented either as the means with standard deviations or as the medians. Continuous variables were compared by using the Student t test, and incidence variables were compared using the c2 test. Differences with P values <.05 were considered statistically significant.
RESULTS Tumor Marker Analysis in HCC Patients Undergoing Liver Transplantation
In part 1 of this study on 663 liver transplantation cases, the clinical features and tumor extents of both groups were very similar to each other (Table 1) [1]. The mean and median pretransplantation AFP levels were 54.5 123.4 ng/mL and 9.8 ng/mL in the HBV group (n ¼ 628) and 65.1 163.7 ng/mL and 21.9 ng/mL in the HCV group (n ¼ 35), respectively. Differences between the mean values did not demonstrate a statistical difference (P ¼ .629), but comparison of the median values did so (P ¼ .011). Pretransplantation AFP levels >20 ng/mL were observed in 267 of 628 HBV patients (42.5%) and 21 of 35 HCV patients (60%) (P ¼ .042). The mean and median pretransplantation PIVKA-II levels were 57.8 114.6 mAU/mL and 23.0 mAU/mL in the HBV group and 79.8 115.1 mAU/mL and 28.0 mAU/mL in the Table 1. Profiles of 663 Hepatocellular Carcinoma Patients Who Have Undergone Liver Transplantation Parameter
Age (y) Male Underlying liver disease Viral hepatitis B Viral hepatitis C Model for End-Stage Liver Disease score Graft-recipient weight ratio Explant hepatocellular carcinoma pathology Maximal tumor size (cm) Tumor number Single (n, %) Multiple (n, %) Within Milan criteria (n, %) Within Asan criteria (n, %)
47.8 6.8 534 (80.5%) 628 (94.7%) 35 (5.3%) 16.4 7.9 0.98 0.18 2.8 1.8 375 288 504 577
(56.6%) (43.4%) (76.0%) (87.0%)
Table 2. Profiles of 2709 Hepatocellular Carcinoma Patients Who Have Undergone Liver Resection Parameter
Age (y) Male Underlying liver disease Viral hepatitis B Viral hepatitis C Nonviral Indocyanine green retention rate at 15 min (%) Explant hepatocellular carcinoma pathology Maximal tumor size (cm) Tumor number Single (n, %) Multiple (n, %) Systematic resection (n, %)
51.6 6.5 2094 (77.3%) 2258 (83.3%) 143 (5.3%) 308 (11.4%) 9.8 4.3 4.8 1.9 2419 (89.3%) 290 (10.7%) 2460 (90.8%)
HCV group, respectively. Comparison of the means did not demonstrate a statistical difference (P ¼ .269), but comparison of the medians did so (P ¼ .043). Pretransplantation PIVKA-II levels >40 mAU/mL were observed in 192 of 628 HBV patients (30.6%) and 15 of 35 HCV patients (42.9%) (P ¼ .127). Distribution of AFP and PIVKA-II values in HBV and HCV groups in the liver transplantation group are shown in Fig 1. Tumor Marker Analysis in HCC Patients Undergoing Liver Resection
Part 2 of this study revealed the preoperative tumor marker status in 2709 patients who had undergone liver resection. The mean and median preoperative AFP levels were 8646.2 58207.6 ng/mL and 23.6 ng/mL in the HBV group (n ¼ 2258) and 929.9 5527.2 ng/mL and 13.4 ng/mL in the HCV group (n ¼ 143), respectively. Differences between the mean values did not demonstrate a statistical difference (P ¼ .113), but comparison of the median values showed a significant difference (P ¼ .006). Preoperative AFP levels >20 ng/mL were observed in 1168 of 2258 HBV patients (51.7%) and 62 of 143 HCV patients (43.3%) (P ¼ .052). The mean and median preoperative PIVKA-II levels were 2132.7 7359.7 ng/mL and 66.0 ng/mL in the HBV group (n ¼ 2258) and 1055.2 4105.4 ng/mL and 49.0 ng/ mL in the HCV group (n ¼ 143), respectively. Differences between the mean values did not demonstrate a statistical difference (P ¼ .083), but comparison of the median values showed a significant difference (P ¼ .045). Pretransplantation AFP levels >20 ng/mL were observed in 1348 of 2258 HBV patients (59.7%) and 81 of 143 HCV patients (56.6%) (P ¼ .47). In the NBNC group (n ¼ 308), the mean and median preoperative AFP levels were 6401.4 37916.3 ng/mL and 6.6 ng/mL, respectively, and mean and median preoperative PIVKA-II levels were 2670.3 7017.6 mAU/mL and 74.5 mAU/mL, respectively. Preoperative AFP levels >20 ng/mL were observed in 110 of 308 NBNC patients (35.7%), and PIVKA-II levels >40 mAU/mL were observed in 188 of 308 NBNC patients (61.0%).
890
HA, HWANG, KIM ET AL
Fig 1. Scatter plots on distribution of AFP and PIVKA-II values in HBV (A) and HCV (B) groups of patients who had undergone liver transplantation. AFP, alpha-fetoprotein; PIVKA-II, proteins induced by vitamin K absence or antagonist-II; HBV, hepatitis B virus; HCV, hepatitis C virus.
Distribution of AFP and PIVKA-II values in the HBV and HCV groups are depicted in Fig 2. The proportions of tumor marker elevation depending on background liver disease are summarized in Table 3.
Receiver-Operator Characteristic (ROC) Curve Analysis
ROC curve analyses were performed for the 2709 patients who have undergone liver resection. In 2258 HBV patients, the AFP cutoff of 20 ng/mL was set and ROC curve analyses
Fig 2. Scatter plots on distribution of AFP and PIVKA-II values in HBV (A), HCV (B), and non-HBV and non-HCV (NBNC) (C) groups of patients who had undergone liver resection. AFP, alpha-fetoprotein; PIVKA-II, proteins induced by vitamin K absence or antagonist-II; HBV, hepatitis B virus; HCV, hepatitis C virus.
HEPATOCELLULAR CARCINOMA TUMOR MARKERS
891
Table 3. The Proportions of Preoperative Serum Tumor Marker Elevation in Hepatocellular Carcinoma Patients Who Have Undergone Liver Transplantation or Liver Resection Liver Transplantation
AFP > 20 ng/mL PIVKA-II > 40 mAU/mL
Liver Resection
HBV (n ¼ 628)
HCV (n ¼ 35)
P Value
HBV (n ¼ 2258)
HCV (n ¼ 143)
NBNC (n ¼ 308)
P Value*
42.5% 30.6%
60.0% 42.9%
.042 .127
51.7% 59.7%
43.3% 56.6%
35.7% 60.1%
.052 .47
Abbreviations: HBV, hepatitis B virus; HCV, hepatitis C virus, NBNC, non-HBV non-HCV; AFP, alpha-fetoprotein; PIVKA-II, proteins induced by vitamin K absence or antagonist-II. *P value between HBV and HCV groups.
Fig 3. Receiver-operator characteristic curve analysis regarding expression of tumor markers. (A) Expression of PIVKA-II at an AFP cutoff of 20 ng/mL in HBV patients; (B) expression of AFP at a PIVKA-II cutoff of 40 mAU/mL in HBV patients; (C) expression of PIVKAII at an AFP cutoff of 20 ng/mL in HCV patients; (D) expression of AFP at PIVKA-II cutoff of 40 mAU/mL in HCV patients; (E) expression of PIVKA-II at an AFP cutoff of 20 ng/mL in non-HBV non-HCV patients; and (F) expression of AFP at a PIVKA-II cutoff of 40 mAU/ mL in non-HBV non-HCV patients. AFP, alpha-fetoprotein; PIVKA-II, proteins induced by vitamin K absence or antagonist-II; HBV, hepatitis B virus; HCV, hepatitis C virus.
892
were performed regarding expression of PIVKA-II (Fig 3A), showing an area under the curve (AUC) of 0.634. The sensitivity and specificity were 68.5% and 49.8% at a PIVKA-II cutoff of 40 mAU/mL, respectively. Next, the PIVKA-II cutoff of 40 mAU/mL was set and ROC curve analyses were performed regarding expression of AFP (Fig 3B), showing an AUC of 0.627. The sensitivity and specificity were 59.2% and 59.6% at an AFP cutoff of 20 ng/mL, respectively. In 143 HCV patients, the AFP cutoff of 20 ng/mL was set and ROC curve analyses were performed regarding expression of PIVKA-II (Fig 3C), showing an AUC of 0.544. The sensitivity and specificity were 54.8% and 42.0% at a PIVKA-II cutoff of 40 mAU/mL, respectively. Next, the PIVKA-II cutoff of 40 mAU/mL was set and ROC curve analyses were performed regarding expression of AFP (Fig 3D), showing an AUC of 0.479. The sensitivity and specificity were 42.2% and 55.0% at an AFP cutoff of 20 ng/mL, respectively. In 308 NBNC patients, the AFP cutoff of 20 ng/mL was set and ROC curve analyses were performed regarding expression of PIVKA-II (Fig 3E), showing an AUC of 0.684. The sensitivity and specificity were 74.8% and 46.7% at a PIVKA-II cutoff of 40 mAU/mL, respectively. Next, the PIVKA-II cutoff of 40 mAU/mL was set and ROC curve analyses were performed regarding expression of AFP (Fig 3F), showing an AUC of 0.612. The sensitivity and specificity were 43.9% and 76.5% at an AFP cutoff of 20 ng/mL, respectively. DISCUSSION
The results of this study suggest that serum AFP and PIVKA-II in HCC patients may be quite variably expressed, not solely depending on background liver diseases. Highly elevated AFP and PIVKA-II levels are often found in HBV patients with advanced HCC. There are several studies that have reported that HCC patients with higher serum PIVKA-II levels demonstrate the distinct clinical features of HCC, such as larger tumor size, more frequent vascular invasion, more intrahepatic metastasis, more frequent recurrence after treatment, and shorter survival [2e6]. However, most of the available studies were conducted on patients with HCV-related HCC, not HBV-related HCC. In addition to our institutional data, we attempted independent analysis of Japanese institutional data (personal communication with Kurume University) in which preoperative AFP level >20 ng/mL was observed in 46.4% of HBV patients and 42.7% of HCV patients (P ¼ .53). PIVKA-II >40 mAU/mL was observed in 63.1% of HBV patients and 56.0% of HCV patients (P ¼ .23). Both AFP and PIVKA-II have complementary roles. However, in this study, ROC curve analyses revealed that the expression pattern of PIVKA-II in patients with elevated AFP was not predictable and vice versa, regardless of underlying liver diseases. The correlation coefficient
HA, HWANG, KIM ET AL
between the levels of AFP and PIVKA-II values in 96 patients with HCC before liver resection was reported to be 0.14, and the sensitivity and specificity were 39% and 87% with an AFP level of 20 ng/mL, 55.2% and 90.9% with a PIVKA-II level of 40 mAU/mL, and 68.8% and 79.2% with a combination of AFP level of 20 ng/mL or PIVKA-II level of 40 mAU/mL, respectively [7]. The results of this study reveal that both AFP and PIVKA-II are variably expressed regardless of the background liver diseases, such as HBV or HCV infection. It is known that the clinical features and mechanisms of carcinogenesis of HCC differ quite markedly depending on the underlying disease. HBV-related HCC more frequently presents as a massive tumor that is more aggressive than HCV-related HCC [8,9]. With regard to carcinogenesis, integrated HBV DNA and HBx proteins have been reported to play a critical role in HBV-related HCC, whereas underlying liver cirrhosis and chronic inflammation of the liver are believed to represent the middle of HCV-associated oncogenesis [10,11]. Considering the differences in oncogenesis and the clinical features of HBVrelated HCC and HCV-related HCC, it may be valuable in the future to evaluate the role of PIVKA-II in patients with HBV-related HCC because PIVKA-II is known to be associated with tumor characteristics and survival in patients with HBV-related HCC [12]. The limitations of this study included that tumor staging was not matched between the HBV and HCV groups and the effects of neoadjuvant therapy were not considered. In conclusion, the results of this study indicate that serum AFP and PIVKA-II may be expressed variably regardless of the type of background liver disease. Further largevolume multicenter studies are needed to evaluate the possibility of the etiology-dependent expression of tumor markers. REFERENCES [1] Lee SG, Hwang S, Moon DB, et al. Expanded indication criteria of living donor liver transplantation for hepatocellular carcinoma at one large-volume center. Liver Transpl 2008;14:935. [2] Tang W, Miki K, Kokudo N, et al. Des-gamma-carboxy prothrombin in cancer and non-cancer liver tissue of patients with hepatocellular carcinoma. Int J Oncol 2003;22:969. [3] Koike Y, Shiratori Y, Sato S, et al. Des-gamma-carboxy prothrombin as a useful predisposing factor for the development of portal venous invasion in patients with hepatocellular carcinoma: a prospective analysis of 227 patients. Cancer 2001;91:561. [4] Nakamura S, Nouso K, Sakaguchi K, et al. Sensitivity and specificity of des-gamma-carboxy prothrombin for diagnosis of patients with hepatocellular carcinomas varies according to tumor size. Am J Gastroenterol 2006;101:2038. [5] Nagaoka S, Yatsuhashi H, Hamada H, et al. The desgammacarboxy prothrombin index is a new prognostic indicator for hepatocellular carcinoma. Cancer 2003;98:2671. [6] Kaibori M, Matsui Y, Yanagida H, et al. Positive status of alpha-fetoprotein and des-gamma-carboxy prothrombin: important prognostic factor for recurrent hepatocellular carcinoma. World J Surg 2004;28:702. [7] Yamamoto K, Imamura H, Matsuyama Y, et al. AFP, AFPL3, DCP, and GP73 as markers for monitoring treatment response
HEPATOCELLULAR CARCINOMA TUMOR MARKERS and recurrence and as surrogate markers of clinicopathological variables of HCC. J Gastroenterol 2010;45:1272. [8] Cantarini MC, Trevisani F, Morselli-Labate AM, et al. Effect of the etiology of viral cirrhosis on the survival of patients with hepatocellular carcinoma. Am J Gastroenterol 2006;101:91. [9] Tanabe G, Nuruki K, Baba Y, et al. A comparison of hepatocellular carcinoma associated with HBV or HCV infection. Hepatogastroenterology 1999;46:2442.
893 [10] Wands JR. Prevention of hepatocellular carcinoma. N Engl J Med 2004;351:1567. [11] Anzola M. Hepatocellular carcinoma. Role of hepatitis B and hepatitis C viruses proteins in hepatocarcinogenesis. J Viral Hepat 2004;11:383. [12] Kim HS, Park JW, Jang JS, et al. Prognostic values of alphafetoprotein and protein induced by vitamin K absence or antagonist-II in hepatitis B virus-related hepatocellular carcinoma: a prospective study. J Clin Gastroenterol 2009;43:482.
I
T IS controversial that different tumor markers are often emphasized between Korea and Japan for the diagnosis and follow-up treatment of hepatocellular carcinoma (HCC) in liver resection patients and liver transplant recipients. In fact, alpha-fetoprotein (AFP) seems to be more frequently emphasized than protein induced by vitamin K absence or antagonist-II (PIVKA-II or des-gamma-carboxy prothrombin [DCP]) in Korea, where hepatitis B virus (HBV) infection is predominant, and vice versa in Japan, where hepatitis C virus (HCV) infection is more prevalent [1e6]. At this time, a question was raised regarding the possibility of a different expression of HCC tumor marker according to the primary diseases of the background livers. This study was conducted to compare the expression 0041-1345/14/$esee front matter http://dx.doi.org/10.1016/j.transproceed.2013.12.037 888
patterns of serum AFP and PIVKA-II in HCC patients undergoing liver transplantation and liver resection at a high-volume single institution.
From the Department of Surgery (S.H., K.H.K., Y.J.L., C.S.A., D.B.M., T.Y.H., G.W.S., K.M.P., S.G.L.) and the Asan Institute for Life Sciences (N.K.), Asan Medical Center, University of Ulsan College of Medicine, Seoul, Korea. Address reprint requests to Shin Hwang, MD, PhD, Division of Hepatobiliary Surgery and Liver Transplantation, Department of Surgery, Asan Medical Center, University of Ulsan College of Medicine, Poongnap-dong, Songpa-gu, Seoul, 138-736, Korea. E-mail: [email protected] ª 2014 by Elsevier Inc. All rights reserved. 360 Park Avenue South, New York, NY 10010-1710 Transplantation Proceedings, 46, 888e893 (2014)
HEPATOCELLULAR CARCINOMA TUMOR MARKERS
889
PATIENTS AND METHODS In part 1 of this study, using our institution’s liver transplantation database containing more than 3000 adult cases, 663 recipients who had been diagnosed with viral hepatitis and HCC were selected. To objectively compare the patient parameters, patients with AFP <1000 ng/mL and PIVKA-II <1000 mAU/mL were included (Table 1). They were divided into the HBV group (n ¼ 628) and the HCV group (n ¼ 35), and their medical records were retrospectively reviewed. In part 2 of this study, a cohort of 2709 patients who underwent HCC resection, which consisted of consecutive 2258 HBV patients, 143 HCV patients, and 308 non-HBV non-HCV (NBNC) patients, had their AFP and PIVKA-II expression levels analyzed (Table 2). If both AFP and PIVKA-II were not measured at the same time, such patients were excluded from this study. No value limitation for exclusion was applied to the HCC resection group. Our study protocol was approved by the institutional review board of our institution. All numerical data are presented either as the means with standard deviations or as the medians. Continuous variables were compared by using the Student t test, and incidence variables were compared using the c2 test. Differences with P values <.05 were considered statistically significant.
RESULTS Tumor Marker Analysis in HCC Patients Undergoing Liver Transplantation
In part 1 of this study on 663 liver transplantation cases, the clinical features and tumor extents of both groups were very similar to each other (Table 1) [1]. The mean and median pretransplantation AFP levels were 54.5 123.4 ng/mL and 9.8 ng/mL in the HBV group (n ¼ 628) and 65.1 163.7 ng/mL and 21.9 ng/mL in the HCV group (n ¼ 35), respectively. Differences between the mean values did not demonstrate a statistical difference (P ¼ .629), but comparison of the median values did so (P ¼ .011). Pretransplantation AFP levels >20 ng/mL were observed in 267 of 628 HBV patients (42.5%) and 21 of 35 HCV patients (60%) (P ¼ .042). The mean and median pretransplantation PIVKA-II levels were 57.8 114.6 mAU/mL and 23.0 mAU/mL in the HBV group and 79.8 115.1 mAU/mL and 28.0 mAU/mL in the Table 1. Profiles of 663 Hepatocellular Carcinoma Patients Who Have Undergone Liver Transplantation Parameter
Age (y) Male Underlying liver disease Viral hepatitis B Viral hepatitis C Model for End-Stage Liver Disease score Graft-recipient weight ratio Explant hepatocellular carcinoma pathology Maximal tumor size (cm) Tumor number Single (n, %) Multiple (n, %) Within Milan criteria (n, %) Within Asan criteria (n, %)
47.8 6.8 534 (80.5%) 628 (94.7%) 35 (5.3%) 16.4 7.9 0.98 0.18 2.8 1.8 375 288 504 577
(56.6%) (43.4%) (76.0%) (87.0%)
Table 2. Profiles of 2709 Hepatocellular Carcinoma Patients Who Have Undergone Liver Resection Parameter
Age (y) Male Underlying liver disease Viral hepatitis B Viral hepatitis C Nonviral Indocyanine green retention rate at 15 min (%) Explant hepatocellular carcinoma pathology Maximal tumor size (cm) Tumor number Single (n, %) Multiple (n, %) Systematic resection (n, %)
51.6 6.5 2094 (77.3%) 2258 (83.3%) 143 (5.3%) 308 (11.4%) 9.8 4.3 4.8 1.9 2419 (89.3%) 290 (10.7%) 2460 (90.8%)
HCV group, respectively. Comparison of the means did not demonstrate a statistical difference (P ¼ .269), but comparison of the medians did so (P ¼ .043). Pretransplantation PIVKA-II levels >40 mAU/mL were observed in 192 of 628 HBV patients (30.6%) and 15 of 35 HCV patients (42.9%) (P ¼ .127). Distribution of AFP and PIVKA-II values in HBV and HCV groups in the liver transplantation group are shown in Fig 1. Tumor Marker Analysis in HCC Patients Undergoing Liver Resection
Part 2 of this study revealed the preoperative tumor marker status in 2709 patients who had undergone liver resection. The mean and median preoperative AFP levels were 8646.2 58207.6 ng/mL and 23.6 ng/mL in the HBV group (n ¼ 2258) and 929.9 5527.2 ng/mL and 13.4 ng/mL in the HCV group (n ¼ 143), respectively. Differences between the mean values did not demonstrate a statistical difference (P ¼ .113), but comparison of the median values showed a significant difference (P ¼ .006). Preoperative AFP levels >20 ng/mL were observed in 1168 of 2258 HBV patients (51.7%) and 62 of 143 HCV patients (43.3%) (P ¼ .052). The mean and median preoperative PIVKA-II levels were 2132.7 7359.7 ng/mL and 66.0 ng/mL in the HBV group (n ¼ 2258) and 1055.2 4105.4 ng/mL and 49.0 ng/ mL in the HCV group (n ¼ 143), respectively. Differences between the mean values did not demonstrate a statistical difference (P ¼ .083), but comparison of the median values showed a significant difference (P ¼ .045). Pretransplantation AFP levels >20 ng/mL were observed in 1348 of 2258 HBV patients (59.7%) and 81 of 143 HCV patients (56.6%) (P ¼ .47). In the NBNC group (n ¼ 308), the mean and median preoperative AFP levels were 6401.4 37916.3 ng/mL and 6.6 ng/mL, respectively, and mean and median preoperative PIVKA-II levels were 2670.3 7017.6 mAU/mL and 74.5 mAU/mL, respectively. Preoperative AFP levels >20 ng/mL were observed in 110 of 308 NBNC patients (35.7%), and PIVKA-II levels >40 mAU/mL were observed in 188 of 308 NBNC patients (61.0%).
890
HA, HWANG, KIM ET AL
Fig 1. Scatter plots on distribution of AFP and PIVKA-II values in HBV (A) and HCV (B) groups of patients who had undergone liver transplantation. AFP, alpha-fetoprotein; PIVKA-II, proteins induced by vitamin K absence or antagonist-II; HBV, hepatitis B virus; HCV, hepatitis C virus.
Distribution of AFP and PIVKA-II values in the HBV and HCV groups are depicted in Fig 2. The proportions of tumor marker elevation depending on background liver disease are summarized in Table 3.
Receiver-Operator Characteristic (ROC) Curve Analysis
ROC curve analyses were performed for the 2709 patients who have undergone liver resection. In 2258 HBV patients, the AFP cutoff of 20 ng/mL was set and ROC curve analyses
Fig 2. Scatter plots on distribution of AFP and PIVKA-II values in HBV (A), HCV (B), and non-HBV and non-HCV (NBNC) (C) groups of patients who had undergone liver resection. AFP, alpha-fetoprotein; PIVKA-II, proteins induced by vitamin K absence or antagonist-II; HBV, hepatitis B virus; HCV, hepatitis C virus.
HEPATOCELLULAR CARCINOMA TUMOR MARKERS
891
Table 3. The Proportions of Preoperative Serum Tumor Marker Elevation in Hepatocellular Carcinoma Patients Who Have Undergone Liver Transplantation or Liver Resection Liver Transplantation
AFP > 20 ng/mL PIVKA-II > 40 mAU/mL
Liver Resection
HBV (n ¼ 628)
HCV (n ¼ 35)
P Value
HBV (n ¼ 2258)
HCV (n ¼ 143)
NBNC (n ¼ 308)
P Value*
42.5% 30.6%
60.0% 42.9%
.042 .127
51.7% 59.7%
43.3% 56.6%
35.7% 60.1%
.052 .47
Abbreviations: HBV, hepatitis B virus; HCV, hepatitis C virus, NBNC, non-HBV non-HCV; AFP, alpha-fetoprotein; PIVKA-II, proteins induced by vitamin K absence or antagonist-II. *P value between HBV and HCV groups.
Fig 3. Receiver-operator characteristic curve analysis regarding expression of tumor markers. (A) Expression of PIVKA-II at an AFP cutoff of 20 ng/mL in HBV patients; (B) expression of AFP at a PIVKA-II cutoff of 40 mAU/mL in HBV patients; (C) expression of PIVKAII at an AFP cutoff of 20 ng/mL in HCV patients; (D) expression of AFP at PIVKA-II cutoff of 40 mAU/mL in HCV patients; (E) expression of PIVKA-II at an AFP cutoff of 20 ng/mL in non-HBV non-HCV patients; and (F) expression of AFP at a PIVKA-II cutoff of 40 mAU/ mL in non-HBV non-HCV patients. AFP, alpha-fetoprotein; PIVKA-II, proteins induced by vitamin K absence or antagonist-II; HBV, hepatitis B virus; HCV, hepatitis C virus.
892
were performed regarding expression of PIVKA-II (Fig 3A), showing an area under the curve (AUC) of 0.634. The sensitivity and specificity were 68.5% and 49.8% at a PIVKA-II cutoff of 40 mAU/mL, respectively. Next, the PIVKA-II cutoff of 40 mAU/mL was set and ROC curve analyses were performed regarding expression of AFP (Fig 3B), showing an AUC of 0.627. The sensitivity and specificity were 59.2% and 59.6% at an AFP cutoff of 20 ng/mL, respectively. In 143 HCV patients, the AFP cutoff of 20 ng/mL was set and ROC curve analyses were performed regarding expression of PIVKA-II (Fig 3C), showing an AUC of 0.544. The sensitivity and specificity were 54.8% and 42.0% at a PIVKA-II cutoff of 40 mAU/mL, respectively. Next, the PIVKA-II cutoff of 40 mAU/mL was set and ROC curve analyses were performed regarding expression of AFP (Fig 3D), showing an AUC of 0.479. The sensitivity and specificity were 42.2% and 55.0% at an AFP cutoff of 20 ng/mL, respectively. In 308 NBNC patients, the AFP cutoff of 20 ng/mL was set and ROC curve analyses were performed regarding expression of PIVKA-II (Fig 3E), showing an AUC of 0.684. The sensitivity and specificity were 74.8% and 46.7% at a PIVKA-II cutoff of 40 mAU/mL, respectively. Next, the PIVKA-II cutoff of 40 mAU/mL was set and ROC curve analyses were performed regarding expression of AFP (Fig 3F), showing an AUC of 0.612. The sensitivity and specificity were 43.9% and 76.5% at an AFP cutoff of 20 ng/mL, respectively. DISCUSSION
The results of this study suggest that serum AFP and PIVKA-II in HCC patients may be quite variably expressed, not solely depending on background liver diseases. Highly elevated AFP and PIVKA-II levels are often found in HBV patients with advanced HCC. There are several studies that have reported that HCC patients with higher serum PIVKA-II levels demonstrate the distinct clinical features of HCC, such as larger tumor size, more frequent vascular invasion, more intrahepatic metastasis, more frequent recurrence after treatment, and shorter survival [2e6]. However, most of the available studies were conducted on patients with HCV-related HCC, not HBV-related HCC. In addition to our institutional data, we attempted independent analysis of Japanese institutional data (personal communication with Kurume University) in which preoperative AFP level >20 ng/mL was observed in 46.4% of HBV patients and 42.7% of HCV patients (P ¼ .53). PIVKA-II >40 mAU/mL was observed in 63.1% of HBV patients and 56.0% of HCV patients (P ¼ .23). Both AFP and PIVKA-II have complementary roles. However, in this study, ROC curve analyses revealed that the expression pattern of PIVKA-II in patients with elevated AFP was not predictable and vice versa, regardless of underlying liver diseases. The correlation coefficient
HA, HWANG, KIM ET AL
between the levels of AFP and PIVKA-II values in 96 patients with HCC before liver resection was reported to be 0.14, and the sensitivity and specificity were 39% and 87% with an AFP level of 20 ng/mL, 55.2% and 90.9% with a PIVKA-II level of 40 mAU/mL, and 68.8% and 79.2% with a combination of AFP level of 20 ng/mL or PIVKA-II level of 40 mAU/mL, respectively [7]. The results of this study reveal that both AFP and PIVKA-II are variably expressed regardless of the background liver diseases, such as HBV or HCV infection. It is known that the clinical features and mechanisms of carcinogenesis of HCC differ quite markedly depending on the underlying disease. HBV-related HCC more frequently presents as a massive tumor that is more aggressive than HCV-related HCC [8,9]. With regard to carcinogenesis, integrated HBV DNA and HBx proteins have been reported to play a critical role in HBV-related HCC, whereas underlying liver cirrhosis and chronic inflammation of the liver are believed to represent the middle of HCV-associated oncogenesis [10,11]. Considering the differences in oncogenesis and the clinical features of HBVrelated HCC and HCV-related HCC, it may be valuable in the future to evaluate the role of PIVKA-II in patients with HBV-related HCC because PIVKA-II is known to be associated with tumor characteristics and survival in patients with HBV-related HCC [12]. The limitations of this study included that tumor staging was not matched between the HBV and HCV groups and the effects of neoadjuvant therapy were not considered. In conclusion, the results of this study indicate that serum AFP and PIVKA-II may be expressed variably regardless of the type of background liver disease. Further largevolume multicenter studies are needed to evaluate the possibility of the etiology-dependent expression of tumor markers. REFERENCES [1] Lee SG, Hwang S, Moon DB, et al. Expanded indication criteria of living donor liver transplantation for hepatocellular carcinoma at one large-volume center. Liver Transpl 2008;14:935. [2] Tang W, Miki K, Kokudo N, et al. Des-gamma-carboxy prothrombin in cancer and non-cancer liver tissue of patients with hepatocellular carcinoma. Int J Oncol 2003;22:969. [3] Koike Y, Shiratori Y, Sato S, et al. Des-gamma-carboxy prothrombin as a useful predisposing factor for the development of portal venous invasion in patients with hepatocellular carcinoma: a prospective analysis of 227 patients. Cancer 2001;91:561. [4] Nakamura S, Nouso K, Sakaguchi K, et al. Sensitivity and specificity of des-gamma-carboxy prothrombin for diagnosis of patients with hepatocellular carcinomas varies according to tumor size. Am J Gastroenterol 2006;101:2038. [5] Nagaoka S, Yatsuhashi H, Hamada H, et al. The desgammacarboxy prothrombin index is a new prognostic indicator for hepatocellular carcinoma. Cancer 2003;98:2671. [6] Kaibori M, Matsui Y, Yanagida H, et al. Positive status of alpha-fetoprotein and des-gamma-carboxy prothrombin: important prognostic factor for recurrent hepatocellular carcinoma. World J Surg 2004;28:702. [7] Yamamoto K, Imamura H, Matsuyama Y, et al. AFP, AFPL3, DCP, and GP73 as markers for monitoring treatment response
HEPATOCELLULAR CARCINOMA TUMOR MARKERS and recurrence and as surrogate markers of clinicopathological variables of HCC. J Gastroenterol 2010;45:1272. [8] Cantarini MC, Trevisani F, Morselli-Labate AM, et al. Effect of the etiology of viral cirrhosis on the survival of patients with hepatocellular carcinoma. Am J Gastroenterol 2006;101:91. [9] Tanabe G, Nuruki K, Baba Y, et al. A comparison of hepatocellular carcinoma associated with HBV or HCV infection. Hepatogastroenterology 1999;46:2442.
893 [10] Wands JR. Prevention of hepatocellular carcinoma. N Engl J Med 2004;351:1567. [11] Anzola M. Hepatocellular carcinoma. Role of hepatitis B and hepatitis C viruses proteins in hepatocarcinogenesis. J Viral Hepat 2004;11:383. [12] Kim HS, Park JW, Jang JS, et al. Prognostic values of alphafetoprotein and protein induced by vitamin K absence or antagonist-II in hepatitis B virus-related hepatocellular carcinoma: a prospective study. J Clin Gastroenterol 2009;43:482.